cout behaves weirdly - prints elements last-to-first

[pmitas]

As an assignment for my beginner C++ course I'm writing a stack that holds strings (char tables). It has an overloaded "operator const char*()" that should enable me to use it's pop method in couts. Most of the time it seems to work nicely, but if I try to pop more than one element in a single line, like this: cout<<s.pop()<<endl<<s.pop<<endl;, it prints the values in reverse order. If I use two couts in two lines, it works well.

For example the code:

int main()
{
	stack s;
	s.push("1");
	s.push("2");
	cout<<s.pop()<<endl<<s.pop()<<endl<<endl;
	s.push("1");
	s.push("2");
	cout<<s.pop()<<endl;
	cout<<s.pop()<<endl;
	return 0;
}

will result in the following output:

1
2

2
1

Here's the source, sorry for the Polish labels, I'll include a little dictionary in the comments ;P

stack.h:

#include <iostream>

class element
{
	public:
		element() : next ( 0 ), tab ( 0 ) {};
		element ( element& );
		~element();
		element ( const char*, element* );
		element ( const char* );
		const char* dajtablice()
		{
			return tab;
		}
		element& operator= ( const element& );
		void wpisz();  //unimportant
		friend class stos;
		operator const char*();
	private:
		element* next;
		char* tab;
};


class stos //"stack"
{
	public:
		stos() :pierwszy ( 0 ){};
		stos ( const stos& );
		~stos();
		void dodaj ( const char* ); //"push"
		element zdejmij(); // "pop"
		element& pokaz(); //unimportant
	private:
		element* pierwszy; //"first"
};

stack.cpp:

#include "stos.h"
#include <cstring>
using namespace std;
const int MAX=100;

void element::wpisz() //unimportant
{
	char c[MAX];
	cout<<"Napisz se coś"<<endl;
	cin>>c;
	tab=new char[strlen ( c ) +1];
	strcpy ( tab,c );
};

element::~element()
{
	if ( tab )
		delete[] tab;
}

element::element ( element& e )
{
	if ( e.tab )
	{
		next=0;
		tab=new char[strlen ( e.tab ) +1];
		strcpy ( tab, e.tab );
	}
}

element& element::operator= ( const element& r )
{
	if ( !r.tab )
	{
		if ( tab )
			delete[] tab;
		tab=0;
		return *this;
	}
	else if ( tab )
	{
		if ( strlen ( tab ) <strlen ( r.tab ) )
		{
			delete[] tab;
			tab=new char[strlen ( r.tab ) +1];
		}
	}
	else
		tab=new char[strlen ( r.tab ) +1];
	strcpy ( tab, r.tab );
	return *this;
}

void stos::dodaj ( const char* c ) //"push"
{
	if ( strlen ( c ) )
	{
		element* drugi=pierwszy;
		pierwszy=new element ( c, drugi );
	}
}

element::element ( const char* c, element* e )
{
	tab=new char[strlen ( c ) +1];
	next=e;
	strcpy ( tab,c );
}

element stos::zdejmij() //"pop"
{
	element ret ( *pierwszy );
	element* temp=pierwszy;
	pierwszy=pierwszy->next;
	delete temp;
	return ret;
}
element& stos::pokaz()
{
	return *pierwszy;
}

element::operator const char*()
{
	return tab;
}

element::element ( const char* c )
{
	next=0;
	tab=new char[strlen ( c ) +1];
	strcpy ( tab,c );
}

stos::~stos()
{
	element* temp;
	while ( pierwszy )
	{
		temp=pierwszy;
		pierwszy=pierwszy->next;
		delete temp;
	}
}

example main.cpp to demonstrate problem:

#include "stos.h"

using namespace std;

int main()
{
	stos s;
	s.dodaj("1");
	s.dodaj("2");
	cout<<s.zdejmij()<<endl<<s.zdejmij()<<endl<<endl;
	s.dodaj("1");
	s.dodaj("2");
	cout<<s.zdejmij()<<endl;
	cout<<s.zdejmij()<<endl;
	return 0;
}

The output, as I said earlier, is:

1
2

2
1

I'd like to know what the hell does this ;) and how can I fix it.
thanks in advance for the help.

[firedraco]

It is because of sequence points (go look it up on wikipedia if you don't know what that is).

The order of function calls on line 10 is undefined, for VC++, it seems to go right to left, so it will call the right zdejmij() first, giving it 2, then call the first one, giving it the value 1. However, like I said, it is undefined, so the program could do it in any order it feels like.

[Bazzy]

The << doesn't guarantee that the operands are evaluated from left to right so the second pop may be called before the first one and as a result you'll get the values to be printed in reverse order

[pmitas]

Thanks for the quick reply ;)